Capstone A

Using Hospital Bed Capacity Prediction During COVID-19 to Determine Feature Importance

Abstract. The Covid-19 pandemic has led to the generation of multiple types of models and feature selection methods in the field of Machine Learning. Since there has been rapid change and new regulations being introduced during the pandemic, modeling and feature selection methods have become increasingly complicated. The purpose of this study is to investigate and dive into key features to help create an understanding for the public and help show preventive measures. This study focuses on the exploration of feature selection though building multiple models, one simple linear model, one more complex model and an average of the two for prediction on impatient hospitalization rates.

Authors:


Missing Values


codebook https://github.com/OxCGRT/covid-policy-tracker/blob/master/documentation/codebook.md

Missing Index Values


***Done Imputing Oxford Dataset***



Import 2nd Dataset


End of Dataset 2 Impute


Start Dataset Merge

CDC VACCINATION DATA

vaccines.loc[2:]


EDA



Graph to compare Cases and Deaths over time

--


Model Building and Evaluation

Data Prep


Modeling

Linear Regression


RandomForestRegressor


Custom Grid Search by Justin Ehly


Model Stats


Sliding Window for 12 months